JP2528386B2 - Failure diagnosis method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a failure diagnosis method used for operation support of factory equipment and the like.

[Conventional technology]

2. Description of the Related Art Conventionally, in factory equipment operated by automatic control such as transfer press processing equipment, an expert system is being adopted for various operation support including failure diagnosis.

This expert system is operated on a single OS such as MS-DOS in many cases in consideration of required capabilities.

Next, a conventional fault diagnosis method using an expert system for driving support will be described.

For example, in the case of the transfer press processing equipment, the state of each place such as press equipment of the equipment is a sequencer [programmable logic controller (PLC)] as a monitoring unit.
Is constantly monitored by.

Then, when some kind of failure occurs in the equipment, the sequencer detects each abnormality occurrence point based on this failure.

Furthermore, the computer of the expert system always operates in the failure occurrence detection mode, and the monitoring output of the sequencer is periodically fetched by the computer to determine whether a failure has occurred.
Nothing is detected.

Then, when some kind of failure occurs and the failure occurrence is detected, the operation mode of the computer shifts from the failure occurrence detection to the diagnostic processing, the failure occurrence detection is stopped, and the failure cause is identified.

This identification is based on the inference based on the information of the abnormal point obtained by the fault detection as the first inference by the diagnostic processing of the forward inference of the production rule configuration of the "if (condition) then (conclusion)" format, It is repeated by determining the next inference from the inference result and firing.

At this time, in order to obtain information necessary for inference, an inquiry screen is displayed immediately before each inference and an information input request for inquiry is issued.

And, according to the issued request, the air pressure of each place,
The following inference is executed every time the current information such as hydraulic pressure is input by acquisition.

Further, when the cause of failure is identified and detected, for example, the cause of failure and a method for recovering from it are displayed on the screen and notified.

[Problems to be Solved by the Invention]

In the case of the conventional failure diagnosis method, when the occurrence of a failure is detected and the mode of the diagnosis processing is entered, thereafter, the inference is repeated until the cause of the failure is identified and detected.

Therefore, even if the cause of the failure is resolved by operating the switch or valve during the diagnosis process and the failure is recovered, the inference for identifying the cause of the failure is continuously executed and is not stopped, and appropriate processing according to the situation can be performed. There is no problem.

Even when the expert system for support is operated on a multi-OS, if it shifts to the diagnostic processing mode as described above, it will be fixed to the diagnostic processing mode unless a recovery operation is performed thereafter. Therefore, the single OS
The same problems as in the case of operating above occur.

It is an object of the present invention to provide a failure diagnosis method that performs appropriate processing according to the situation by a method that can be applied to the failure diagnosis of an expert system configuration operated on a single OS.

[Means for solving the problem]

In order to achieve the above-mentioned object, in the failure diagnosis method of the present invention, after shifting to diagnostic processing by detecting the occurrence of a failure, the monitoring result of the monitoring unit for each free time waiting for input based on the information input request for inquiry. The continuation and elimination of the fault are detected from the above, and only when the elimination of the fault is detected, the diagnosis process is stopped and the process returns to the fault occurrence detection.

[Action]

In the case of the failure diagnosis method of the present invention configured as described above, during diagnosis processing, each time an inquiry information input request is issued, the idle time until information is input based on this request is used. Continuation and elimination of failures are detected.

Then, at the time of recovery from the failure in which the resolution of the failure is detected, the diagnostic process is stopped and the process returns to the failure occurrence detection.

Therefore, during the diagnosis process, the recovery from the failure can be constantly monitored without performing a plurality of processes at the same time, and the failure occurrence detection can be immediately returned at the time of the recovery, and an appropriate process according to the situation can be performed.

〔Example〕

One embodiment will be described with reference to FIGS. 1 to 5.

In FIG. 2 showing the device configuration, 1 is a sequencer (PLC) as a monitoring unit, 2 is a communication unit such as an optical modem interposed between the sequencer 1 and the personal computer 3,
Numerals 4 and 5 are an optimal failure selection section and a failure diagnosis section of the software configuration inside the computer 3, and 6 and 7 are an input section and display section of the computer 3.

Then, the sequencer 1 operates independently of the computer 3 and constantly monitors the state of each part of, for example, a transfer press processing facility as a diagnosis target.

Further, the computer 3 operates on a single OS, and constructs an expert system for support (for failure diagnosis) described below.

As shown in FIG. 1, this system enters a failure occurrence detection mode immediately after the start initial is completed, and executes main processing, communication processing, and internal abnormality check processing programs in sequence at intervals of, for example, approximately 800 μsec. Each process is performed in parallel.

Then, the computer 3 accesses the sequencer 1 via the communication unit 2 by the communication processing, and periodically takes in the monitoring output of the sequencer 1.

Furthermore, the status of each place is grasped by the captured monitoring output to detect whether or not a failure has occurred.

On the other hand, screen display processing and the like are executed by the main processing, and the processing result and the like are displayed on the screen of the sequential display unit 7.

In addition, the internal abnormality check processing allows the computer 3
Whether or not an internal abnormality has occurred is detected, and at the time of detection, predetermined abnormality processing is performed.

When any failure occurrence is detected by detecting the presence or absence of the failure occurrence, the detection result is displayed on the display unit 7 and at the same time, the mode of the diagnostic processing is entered.

At this time, when one or more other faults occur in a chain due to the occurrence of a certain fault and a plurality of faults are simultaneously detected, priority is given to manual selection of which fault is the fundamental fault. In order to improve the inference efficiency by automatically determining the cause of the failure and prioritizing the identification of the cause of the failure, in this embodiment, the selecting unit 4 first selects the failure for forward estimation of the production rule configuration described below. .

Now, for example, regarding a pneumatic system, the expected failure is A,
It is assumed that there are three types, B and C.

In this case, as shown in FIG. 3, the selection unit 4 determines each judgment rule r ₁ , of the following basic form [if (condition) then (conclusion)].
Hold rn in advance.

“If the failure A or the failure C is manually selected by the automatic determination that prioritizes the failure A, the failure A is determined.” Then, the detected failure and the manual selection from the input unit 6 are selected. Based on this, each determination rule of the selection unit 4 is selectively fired according to the immediately preceding inference result, and the root failure is preferentially selected as the optimum failure and automatically determined.

Further, the determination result of the selecting unit 4 is given to the failure diagnosing unit 5, and the diagnosing unit 5 executes the diagnostic process of the forward inference with the production rule configuration described below.

That is, in the diagnosis unit 5, as shown in FIG. 4, the inference unit 5a holds in advance various diagnosis rules R ₁ , ..., Rn having the same format as the determination rule of the selection unit 4.

Further, based on the determination of the optimum failure given from the selection unit 4 and the immediately preceding inference result, a request for inputting the current information necessary for the next inference is made. Therefore, an information input request for inquiry is issued to the display unit 7. Displayed as an interview screen.

Based on the display of this inquiry screen, when the current information such as air pressure and hydraulic pressure at each place is manually read and given from the input unit 6, the input current information (fact) and the knowledge base unit 5b (fact) , Rule of thumb) and the execution result of the selected diagnostic rule, the selected and fired inference is executed.

Based on the result of this inference, the next inference is selected and fired, and the next information input request for inquiry required for this inference is issued.

After that, as long as a mode change based on the situation monitoring described later does not occur, each inference is alternatively selected and sequentially executed, and by this sequential execution, for example, the inference of the contradiction between the normal state and the current state is repeated, As shown in FIG. 5, with respect to the selected fault B, the intermediate facts of each stage are selected as the starting point, and the cause of the failure is finally identified. It is diagnosed (detected).

Then, when the cause of the failure is diagnosed, the cause and the method for recovering from the cause are displayed on the screen of the display unit 7.

When the failure is recovered by the recovery operation based on this screen display, the diagnostic processing mode is automatically terminated, and the subsequent restart operation or the like causes the processing from the start in FIG. 1 to be performed again.

By the way, if the cause of the failure is resolved and the system is restored after shifting to the diagnostic processing mode, the subsequent inference is useless, and it is desirable to immediately return to the failure occurrence detection mode.

Then, it is possible to perform the processing of the failure occurrence detection mode at the same time during the processing of the diagnostic processing mode, but especially in the case of an expert system operated on a single OS, the processing of both modes cannot be performed at the same time.

Therefore, in this embodiment, as shown in FIG. 1, every time an information input request is made by displaying an inquiry screen,
The computer 5 monitors the status of the communication process and the return determination process by using the free time waiting for the input until the information is input according to this request.

Then, the latest monitoring output of the sequencer 1 is taken in by the communication processing in the same manner as in the case of detecting the failure occurrence, and the continuation and elimination of the failure are constantly monitored and detected by the restoration determination processing based on this monitoring output.

Based on this monitoring, the diagnostic process continues while the continuation of the fault is detected, but when the fault is resolved and the fault recovers, the diagnostic process mode is immediately stopped, and the fault detection mode is exited from this mode. Return to mode.

Therefore, after shifting to the diagnostic processing mode, if the system recovers from a failure by constant monitoring using the free time waiting for input based on the information input request, it immediately returns to the failure occurrence detection mode, and appropriate processing according to the situation can be performed. Done.

In the case of the above-mentioned embodiment, since the optimum failure is selected by the selection unit 4 before the estimation of the diagnostic processing is performed, there is an advantage that the failure cause can be identified very efficiently.

And, of course, it can be applied even when the optimum failure is not selected.

Further, it goes without saying that the diagnosis target may be various factory equipment other than the transfer press processing equipment.

〔The invention's effect〕

Since the present invention is configured as described above,
The following effects are achieved.

After shifting to the diagnostic process by detecting the occurrence of a failure, the free time waiting for input based on the information input request for inquiry is used, and the continuation and elimination of the failure are detected from the latest monitoring output of the monitoring unit for each free time. Constantly monitors recovery from failures,
Since the diagnosis process is stopped at the time of recovery and the failure occurrence detection is immediately returned to, the method that can also be applied to the failure diagnosis of the expert system configuration for support operated on the single OS performs appropriate processing according to the situation. be able to.

[Brief description of drawings]

1 to 5 show one embodiment of the failure diagnosis method of the present invention. FIG. 1 is a flowchart of processing procedure, FIG. 2 is a block diagram, and FIGS. 3, 4, and 5 are respectively. Second
It is explanatory drawing of a partial process of a figure. 1 ... Sequencer, 3 ... Computer, 5 ... Diagnosis section, 6 ... Input section, 7 ... Display section.

Claims (1)

(57) [Claims] 1. A production rule based on detection of a failure occurrence based on a monitoring output of a monitoring unit that constantly monitors the state of each part of the diagnosis object, and whether the failure occurrence of the diagnosis object is detected repeatedly. The process proceeds to the diagnostic process of the forward inference of the configuration, issues the information input request for the inquiry, executes the inference based on the current information input according to the information input request, and fires the next inference. In a failure diagnosis method that repeats until the cause of failure is identified, after the diagnosis processing is performed, continuation and resolution of the failure are detected from the monitoring result of the monitoring unit for each free time waiting for input based on the information input request. A fault diagnosis method, wherein the diagnosis process is stopped and the process returns to the fault occurrence detection only when the elimination of the fault is detected.